Web Services

Erik Sliman writes "Why are all the IT companies suddenly interested in open standards with web services? An OpenStandards.net article explores the issues surrounding the somewhat vague term."

Becuase of Stupidity of course

[Telastyn]

It's becoming more and more common that the "Internet" is just Internet Explorer to most people. So some smart fellow thinks it'd be a grand idea if services could be served this way, to appease the lowest common denominator. PHB's get ahold of it, and wham! off it goes to the media, and in 2-3 years everyone (hopefully) realises what a bad idea it was.

If you want a unified 'client' for all services, make one, don't kludge everything onto http. Please...

Re:It's Like Most Bandwagons...

[cbowland]

Funny you mention that. Below is the text of an email I just received from oreilly. I guess you can't hop up on that wagon fast enough (too bad the stepstool is so bloody expensive).

Planning for Web Services is a new report from O'Reilly Research, written by industry visionary Clay Shirky. This report guides CTOs and CIOs through the inflated claims, competing standards, and amalgam of acronyms to arrive at a realistic appraisal of the business impact of Web Services. Topics include how Web Services can replace EDI, who the major players are and what they really offer, as well as the hurdles to implementing Web Services today. A must-read for anyone developing a Web Services business strategy. $495 Save $100! Just use code # wsrelj when ordering by phone [phone] or 707-827-7000) or email ([e-mail]) and you can get this invaluable report for only $395. Offer expires May 10, 2002

We use web services

[Anonymous Coward]

At work we've been using web services to make eligibility requests to insurance company databases. The reason it's nice is

1) no connectivity issues, it's just https over the Net, and

2) no data format issues. In .Net at least, you write a web service just like a local procedure. It has parameters which can be arbitrarily complex objects. Hit a button, it exports WSDL. Send that to your partner, who hits another button, now they can call your web service just like they were calling the procedure locally. No muss, no fuss, and any VB programmer can pick it up in a day and start using it.

Its a good time to ....

[grid geek]

Its a recession. During boom times like the mid 90's companies were too busy dealing with sales and expanding like crazy to deal with demand. Now that most of the competition has died down, no one expects them to post record profits etc it gives people the chance to think about where to go next.

The web is all very well but HTTP et al. have some serious limitations and were never designed for most of the current technology. For example a dial up connection has the same bandwidth of a dedicated line in the 1970's so ASDL/Cable modems etc were never considered.

The reason for all the demand now is the scientific community and all the Grid projects around the world, just because there's a recession doesn't stop them and their data requirements make Google look like a small fry (20TB of data for Google vs 600TB for BaBar at SLAC [stanford.edu]).

The other issue is business - they've all got on the band wagon of internet sales as an extra sales channel so they can grow this, but its not going to be the sudden revenue increase it was initially. Web Services offer the opportunity for companies to increase productivity and efficency which is why the tech companies are investing in it now so when the economy changes and the corporate clients come back they have something new to go on about.

That's what happening

[wiredog]

Well, sort of. The PHB's are realizing that web browsers aren't the best way to do web services/web applications and are looking for a better one.

The problem is that everyone has a web browser. Anything that aims to replace it has to get high distribution at low cost. You want all your customers to have whatever client you use. And it has to be based on a standard so that even if the customers client isn't exactly what you have, it's close enough.

And this is in a world where it can be difficult to get IE and Mozilla to play nicely together.

Web Services = Inherently Insecure

[hndrcks]

Web Services- you take the crackable and exploitable service on port 'X' and advertise it on port 80 or 443. Just as bad, just as exploitable - but now the IT people can't firewall it.

Ummm... hm. Some random thoughts.

[mcc]

This whole thing has me just kind of lost.

The mess of SOAP and RDDI and GESCOM and all these vaguely XML-related, something-to-do-with-port-80 acronyms don't leave me all that impressed; near as i can gather, they're nothing but platforms for people to build platforms on top of, and they won't be of much use until someone takes the foundation of tangled acronyms and builds a common client app that lets you actually use all of these things. I don't take this all this seriously, because knowing the computer industry, i'm pretty sure that by the time "web services" becomes actual services you can use using programs you can download, these services will be using a specific, jury-rigged enough implementation of "web services technology" that you'll be unable to use a given service except with their specific client, and there will be a huge incompatibility rift between MS-based and non-MS-based web services, and basically all of the nice, compatibility-engineering abstractions that the W3C is trying to put together now will be thrown out the window just because the current "web services" standards are so rediculously complicated that no one will be able to come up with an implementation of those standards that really *uses* the full potential of the protocols.

The thing is, though, i really don't care to understand "web services". I understand the following, and i really think it's all i need to know:

I like XML-RPC, because it gives me a really neat, simple way to do simple message passing between programs over the internet, without any more overhead than is absolutely necessary, it's cross-platform and cross-language, it isn't awkward to use in any of the programming languages i've used it in.

XSLT looks really really awesome becuase it's useful and relatively simple, and i really hope we start seeing some tools that can automatically generate some of the XSLT for you, since like all XML tech it's just really verbose.

J2EE looks *interesting*, and all i wish was that it could be interfaced more easily with other languages. I love Jython, but i don't know if i can embrace Java completely until it's possible to let java communicate with arbitrary languages a bit more easily.

Twisted [twistedmatrix.com] looks neat but i don't think i'll ever use it.

I think CORBA would rock my world if it were a bit simpler, or just if someone would find a way to integrate it with the compiler, or just cut out the complicated crap that surrounds using it. C# (whoo, someone's finally figured out that if you make a bunch of languages with the same features but different syntax and macro between them, people will think it's "language cross-compatibility"..) is not the correct way of doing this.

I think "web services" these days comes down mostly to taking the problems with CORBA (it makes stuff simple! but you have to read a 1500 page book before you can start using it!) and putting <html brackets> around them.

I think this article was very interesting, especially the claim that .NET is just microsoft trying to take existing standards and take credit for them. (Although i found it funny that the article gives MS full credit for SOAP. Wasn't the guy who made XML-RPC on the SOAP creation team?)

I would like to know when someone is going to find the balance between J2EE's "everything is nice and fits together and is simple and you just sit down and start doing object oriented programming, but you're chained to the java vm" and the .NET/'web services' "here's a bunch of complicated, bloated standards that take way more bandwidth than they need and that are so abstract you can use them from any language, but also make so many compromises you really don't want to use them unless you're using C# (or a special version of python written for .NET, or a version of C++ that looks exactly like c#..)

You know, it would be really nice if we had *real*, good, turing complete macro languages built into the popular programming languages. Maybe then we wouldn't have to take the C# route of rewriting the compiler just because you want to make it possible to declare a method a "web service" using a single keyword.

Microsoft blames hype for .net woes

[terrymr]

Microsoft is blaming industry hype for the general lack of consumer interest in .net services. Their decision to delay the launch of My Services was apparently because of some kind of consumer backlash against over-hyped web services. read the register article [theregister.co.uk]

Shameless Self-Promotion

[cybermage]

Hey Erik, nice ad:

Organization:
Joshua Branch
Erik Sliman
1449 Larchmont Ave., Dn
Lakewood, OH 44107
US
Phone: [phone]
Email: erik(at)joshuabranch.org

Registrar Name: Register.com
Registrar Whois: whois.register.com
Registrar Homepage: http://www.register.com

Domain Name: OPENSTANDARDS.NET

Created on: Fri, Dec 17, 1999
Expires on: Sun, Dec 17, 2006
Record last updated on: Wed, Mar 06, 2002

Administrative Contact:
Joshua Branch
Erik Sliman
1449 Larchmont Ave., Dn
Lakewood, OH 44107
US
Phone: [phone]
Email: erik(at)joshuabranch.org

Technical Contact, Zone Contact:
Register.Com
Domain Registrar
575 8th Avenue - 11th Floor
New York, NY 10018
US
Phone: [phone]
Fax: [phone]
Email: domain-registrar(at)register.com

Domain servers in listed order:

DNS13.REGISTER.COM 209.67.50.208
DNS14.REGISTER.COM 209.67.50.209

Re:Marketing Hype = More $$$

[rylos]

I agree that the hype is heavy for Web Services. However, I do see benefits for using Web Services.

Making Web Services work in a useful way sometimes takes some creativity. Take Google as an example. With the recent release of the Google API, I was able to use PHP and SOAP to access their search results. One of the methods offered through the service is spell checking. By integrating this spell checking with my company's internal search engine, I now have the ability to make search term suggestions to users. This functionality would be very difficult to provide if it had to be created from scratch.

Web Services will NOT work for all things and in all situations, but they WILL work for some things and in some situations. Creativity is the key.

Web services really do matter

[jamesmartinluther]

While there is a great deal of hype surrounding web services, this group of technologies is going to dominate how the internet is used in the next few years.

It has been an ordeal to get web sites to interact usefully without an end-user clicking on a web page. One big problem is trust. An other is protocol. Sites have so many different ways to get information and to submit information. Worse, site administrators have different ideas about how to make various forms of raw data available to others. Exactly where it is to be found is but one stumbling point, much less how it is structured.

With stuctured data in the form of web services readily available, and clear protocols as to the use of a site's structured data, there will be a lot more interaction between sites and developers of sites.

Most importantly, web services will allow users and sites to become more alike and on more equal ground. This is a powerful change that is already upon us in the form of web sites like slashdot.org and early web services like Napster.

It will be interesting to watch

[e2d2]

I think the service model is an interesting one to take note of and watch over the next few years as the major players roll out their solutions. What I'm most interested in seeing is this technology used to deploy applications across an enterprise. I think this is really where this model can shine in the near future. Currently a lot of enterprises are moving their applications to web and internet based architectures because it can decrease costs of deployment. We all know the Heavy Client vs. Web Client argument and I think there are reasons for using both of them in certain situations. Now imagine corporate users having the ability to subscribe to their enterprises applications as needed. Application management can be consolidated into central locations and the cost of deployment can be decreased significantly. I think what we will need to focus on are tools to enable such deployments and the management of said software in a large, sometimes global, environment. Third party developers can still use the same subscription licensing or develop new licenses for such distribution and the company maintains control of their own information and don't have to rely on 24/7 uptime from an outside service. There are hurdles of course but the technology is here and we heading in that general direction.

I think this is where the first applications in this area will be built and used successfully. The same technology used to deploy applications using web services across enterprises can be used to distribute applications to consumers.

My personal opinion is that service based companies don't exacly have the best track record. I bet chances are pretty good that anyone reading this has had a bad experience at one point with a service provider such as the phone, electric, or cable companies. And also people like the idea of owning something. I myself feel like my whole life is a rental sometimes and it bothers me. It's going to take a lot to push users in this direction and the ones that can execute the best will win. But there is no guarantee it will work. It takes more than just a push or shove to generate a new market, but it can be done.

50% Meat, 50 % Filler

[BroMan]

As it stands now Web services are just the next step in the evolution to a platform-neutral, language-neutral distributed software environment. I see alot of M$ bashing going on, which is somewhat misplaced since some of the biggest backers of web services also happen to be M$'s biggest foes (read - IBM). M$ is however, trying to pull off one of their famous hijacks, with .NET. If they succeed it will be at least partly because of their competitors failure to take the initiative.

Getting back to web services though, they can possibly fill a niche in enterprise computing - and that niche is the ever-present, never fully solved question of how to tie together disparate platforms and software applications in a common enterprise environment. CORBA is the oft-quoted answer, but it is expensive to implement, and hard to get right. Wells Fargo has implemented an interesting solution for distributed programming using something they call Model Driven Architecture [ebizq.net].

Looking at getting systems working together from an IT managers perspective, your always looking at the Big Two - time and money. 'How can I get this system working with my current resources in the least amount of time?'. The complement to that is 'How do we maintain and augment this solution once it goes into production without going through birthing pains?'.

The promise that Web Services is making to IT managers is that they will be able to lower their TCO and increase their ROI by cutting down on the number of changes they have to make to existing systems, while at the same time increasing their flexibility in adding new functionality. To others it makes the promise of providing services that can be metered and billed (wasnt that the promise of CORBA, EJB's, insert favorite distributed model here?).

Of course this is all a pipe dream until they solve some big issues, like security. Transaction management is not as important since web services can actually be implemented in any kind of language you want (read - Implement your own damn transactions).

However I think most IT managers will go blue in the face the first time their Fund Transfer web service is hacked because of a weak 56 bit SSL connection ;)

MS Browser wars

[damien_kane]

What became clear in the browser wars was that open standards was the only way to win any war on the Internet. One may be able to truthfully say that Microsoft has won the browser wars.

Microsoft may have won the browser war with Internet Explorer, but it is not because of open standards.
Any web developer will tell you that javascript et.al., although founded on the same basic functions and routines, is quite different from one browser to another.
Microsoft did not win the browser war through open standards, but by bundling it with Windows.

The fact is, people are too ignorant and lazy to download a completely separate browser, even though it may be (and generally is) more secure than IE. Because of this, the majority of the global online community uses Internet Explorer. The reason this has not changed is because companies realized this. They have thus developed their pages using proprietary MS javascript extensions.

Why build a nuclear car when 99% of gas stations sell gasoline?


"Computer games don't affect kids. If Pac-Man Affected us as children, we would spend all of our time running around in darkened rooms eating magic pills and listening to repetetive electronic music..."

Re:Shameless Self-Promotion

[cybermage]

Why do a WHOIS? Erik's name is on the byline.

OpenStandards was /.ed at the time. Also, I think there's are degrees of shamelessness. It's one thing to submit a story you've found someplace and think others might like. It's a bit shameless to submit a story about an article you wrote that another site posted. It's the epitamy of shameless to submit a story you wrote that is posted on a site you own.

I posted the whois information so that everyone could see just how little research /. needed to do to know that the submitted story was just an Ad.

Re:Marketing Hype = More $$$

[mini me]

I can think of many things I'd love to see exposed via web services (or another RPC type protocol).

What if all the functions of Slashdot were avaliable via SOAP? Then anyone could easily write a Slashdot application that looks more like a news reader, or whatever you want. The value of such a program is debatable, but at least it would be an option. How about weather information, traffic reports, interest rates, currency exchange rates, etc.

Interest rates and exchange rates are an excellent example of how web services could be used. If you are writing a financial application you could then always have up to date rates which would require no human input. Then a web service call could then be made to a bank to make the transaction. All automated from a single program. No human interaction needed.

Web services make perfect sense when someone has information that can be useful to someone else that would otherwise not be avaliable. It allows people to build applications that would not be otherwise possible due to lack of information. It gives computers access to the vast knowledge of the internet instead of just humans.

Of course if no one creates any useful web services then all of this technology will go to waste.

Re:CORBA is too heavy & EJB is too RMI/IIOP de

[dwsauder]

SOAP is supported by Microsoft and IBM, Sun, BEA, and so on. That observation alone seems to suggest that SOAP will go much further than either DCOM or CORBA.

A few years back, I used to wonder what the world of distributed computing would be like if Microsoft decided to support CORBA. Maybe with SOAP, we will get a chance to find out.

BTW, I think Microsoft has no choice but to play along with open standards in web services. If they were to choose otherwise to push their own proprietary web services "standards", their proprietary standards would probably be adopted no more than DCOM.